Recent advances in Data Mining (DM) technologies have piqued the interest of all banking communities as a result of recent advancements in Egyptian banks, where the bank has data that necessitates analysis due to its significance, and one of the most important of these data is related to ATMs.. The role of ATM has been maximized recently due to the national trend of digital transformation and the emergence of COVID-19 pandemic since the ATM provides banks’ customers with financial transaction method in public places without any human interaction. Meanwhile, ATMs confront a slew of challenges, including issues with location, cash, and the operating system, all the problems related to each other. This paper aims to suggest using data mining techniques to solve ATM problems and improve ATM services by building a model that helps measuring ATM efficiency by building data warehouse (DW) and comparing three DM algorithms (Decision Tree (CARET), C5.0, and Naive Bayes). The decision tree (C5) has been shown to be a more convenient technique. The data was analyzed using the R programming language, which revealed a variety of issues and require improvement in order to achieve the required quality.